Speaker-Indentification System based on GMM model.
Robust Text-Independent Speaker Identification Using Gaussian Mixture Speaker Models by D.A. Reynolds in Lincoln Lab., MIT, Lexington, MA, USA and R.C. Rose
Mel Frequency Cepstral Coefficient (MFCC) tutorial
Related system test video can be found here
Thanks for Professor Ivan Selesnick's great class and nice to collaborate with Shihong Fang.